october 2015

The Arctic is melting - so what?

Authors:
James A. Screen
Clara Deser
Lantao Sun
Associate editor:
Erin Conlisk

Abstract
The Arctic sea ice over the North Pole is melting - faster than
humans have ever witnessed - as a result of climate change.
How does this affect the weather elsewhere, though? Does
it? We wanted to know if people outside the Polar Circle
would experience more extreme weather as a result of the
melting.
We entered predications for sea ice loss as input into a
climate model. The goal was to estimate the number of
warm, cold, wet, and dry days as a result of the melted
ice. According to the model, in more northern locations, like
Siberia, Scandinavia and northern North America, people
would experience fewer cold days and more warm days. In
addition, Eastern U.S., Central Europe, Siberia, and all of
Asia can expect fewer dry days while the Mediterranean,
Siberia and northern North America will face more wet days.
Lots of changes in store, right? And not just in the Arctic!

Introduction
If Santa does live at the North Pole, he
lives on a big ice “cube” (really a vast
sheet that is typically only a few meters
thick) floating over the Arctic Ocean.
This could become problematic by 2100
when much of this sea ice is expected
to melt out every summer (Figure 1).
(Sea ice over the Arctic is not to be
confused with Antarctica – the snow-
covered continent at the South Pole.)
Historically, Arctic sea ice would be at
Figure 1. Sea ice in March and September in the year 2000 versus 2100
a maximum in March, when cold temperatures cause the
from the Hadley model. The purple circle is the “Arctic” region. The red
surface of the ocean to freeze (Figure 1, left figure) and
outlines are the extent of sea ice under current conditions in March 2000
about 60% of the sea ice would then melt by September
and September 2000. These outlines are preserved in the second and
(Figure 1, third from left). With global warming, more sea fourth figures from the left to show how sea ice is declining in 2100.
ice is melting in the summer than can form in the winter,
potentially creating an ice-free Arctic Ocean (Figure 1, right

Мore free environmental science resources аt: www.ScienceJournalForKids.org 1


column). At which point, Santa, elves, and reindeer would
need to relocate!
But Santa’s troubles are the tip of the iceberg (which is not
made of sea ice but formed when glaciers on land reach
the ocean and chunks fall into the sea). Loss of Arctic sea
ice can feedback on continued climate change through the
albedo effect (see Glossary) and disrupt Arctic wildlife. Polar
bears are the most common example of wildlife dependent
on sea ice. Polar bears stand on the sea ice when they
hunt; without this platform they can’t catch their prey. But
creatures, such as polar bears, at the “top” of the food web
aren’t the only ones that depend on sea ice for food. Sea-
ice phytoplankton are blooming at different times, disrupting
prey availability for crustaceans that are eaten by larger fish,
such as cod (Post et al. 2013).
In this paper, we want to explore how sea ice loss can affect
extreme weather throughout the Northern Hemisphere (any
part of the world north of the Equator). Extreme weather,
which we define as weather (or atmospheric conditions over
short time frames; climate describes weather over longer
time frames) that has been historically uncommon, has
Methods
We used models (representations, often a simplifications,
of a structure or system) to predict how the climate might
change in the coming decades to centuries. Climate
models are incredibly complex. For example, climate
models must describe how water that evaporates in one
region is transferred through the atmosphere to another
region, potentially hundreds of miles away, and falls to the
ground as precipitation. The exact location of evaporation
and precipitation depend on the land surface (are there
mountains? what vegetation is on the landscape? etc.).
We used two well-established climate models – (i) the UK’s
Hadley Centre model (GAM2) and (ii) the US’s National
Center for Atmospheric Research (CAM4) model. These
models have been used previously (as part of the fifth
Intergovernmental Panel on Climate Change, or IPCC,
report) to predict how much sea ice we expect in 2100. We
do not describe the climate models in detail in this paper
because it would take thousands of pages. If you want to
learn more about it, check out Rubenstein 2010.
We then used these sea ice predictions as model input. We
held everything in the model constant except how much sea
october 2015
The Arctic is melting - so what?
Figure 2. A winter service vehicle abandoned and stuck on the roadside in
Chicago during a winter storm. Extreme cold weather can paralyze entire
cities and economies.
important impacts on humans and nature. For example, for
humans, extreme cold requires greater energy use, often
through the use of fossil fuels, to heat our homes and clear
our roads (Figure 2). For wildlife, behavioral changes (such
as finding more food, better shelter, or a warmer location to
move to) to accommodate extreme cold may not be sufficient
and animal species may die off. Very high rainfall can lead to
flooding that impacts human structures and wildlife.
ice is in the Arctic. Our model functions like our experimental
system in which we compare two treatments: the current
sea ice extent versus the reduced sea ice coverage predicted
for 2100.
We expected that by changing the model’s input (describing
the amount of sea ice in the Arctic), the model would predict
weather changes in other parts of the world, particularly in
locations north of the Equator. We were interested in the
following four measures of extreme weather:
1. Warm days: Percent (%) of days when the maximum
temperature is greater than the 90th percentile of
historic weather,
Cold days: Percent (%) of days when the maximum
2.
temperature is less than the 10th percentile of historic
weather,
3. Wet days: Number of day per years when daily
precipitation is greater than 1 cm, and
4. Dry days: Number of day per years when there is no
precipitation.
2

To define warm and cold days, we looked at the historic
weather for a particular location and ordered all observed
temperature values for a particular day of year. The 10th
percentile reflects the temperature value where 10% of all
the temperature values fall below that value. Because only
a few observations are colder than the 10th percentile, we
Results
The amount of sea ice in the Arctic will have different
impacts on extreme weather depending on the location in
the northern hemisphere. In more northern locations, like
Siberia, Scandinavia, and northern North America, decreases
in sea ice will lead to significantly more warm days and fewer
cold days (Figure 3).
Cold days Warm days
Figure 3. Changes in local weather predicted by the model as a result from Arctic sea ice loss.
Discussion
According to the climate model, loss of sea ice will lead to
local warming in the Arctic which then will move to nearby
locations in northern North America, Scandinavia, and
Siberia. This will cause more warm days and fewer cold days.
october 2015
The Arctic is melting - so what?
define it to be a “cold” day. Similarly, the 90th percentile
reflects the temperature value where 90% of all the
temperature values fall below that value. A day that was
historically hotter than most days (i.e. at the 90th percentile)
is thus defined as a “warm” day.
The impact of sea ice loss on precipitation will occur at both
southern and northern locations. Overall, the model predicts
an increase in precipitation. In the Eastern US, central
Europe, Siberia, and all of Asia the number of dry days are
expected to decrease. In Central Asia, the Mediterranean,
Siberia, and northern North America the number of wet days
will increase (Figure 3).
Wet days Dry days
Further, we can expect a lower variance in the temperature
(in other words, the temperature had more similar values
from day to day).
3
 october 2015
The Arctic is melting - so what?

Conclusion
There are three important things that can be done with the
information in this paper. First, work to reduce greenhouse
gas emissions that cause climate change and associated
melting of sea ice – for the sake of the polar bears and Santa!
In particular, try to avoid using fossil fuels for transportation,
use less energy at home and try to consume less “stuff”.
Almost any item you buy likely required fossil fuels to create
and also to transport to the store.
Second, climate change is a global phenomenon: sea ice
loss in one part of the world can affect the weather where
you live. Similarly, the choices you make will affect the
magnitude of this problem across the entire planet.
Third, understand how scientists create and predict future
climate. A lot of people (especially those as old as Santa!)
have trouble understanding what models do and how to
interpret uncertainty. The best scientists agree that the
changes we are observing now are evidence that climate
change is real and already profoundly impacting our planet.
However, uncertainties about specific climate change
phenomenon (e.g. was this particular hurricane caused
by climate change) are used to justify political inaction.
While you don’t have to be an expert to interpret scientific
information, you do have to practice. Flexing your critical
thinking “muscles” will keep them “strong”.

Discussion Questions
1 Does melting sea ice in the Arctic increase sea level directly? Why or why not?! How would melting over
Antarctica be different?

2 Describe (or draw) the albedo feedback discussed in the paper. (Hint: How does melting sea ice speed up
warming in the Arctic?)

3 According to the climate model experiments, what types of weather extremes (cold, warm, wet, dry) are
expected to change in number where you live? Think of one way that this change would be a good thing
and one way it would be a bad thing.

REFERENCES
Screen JA, Deser C, Sun L. 2015. Projected changes in regional climate extremes arising from Arctic sea ice loss.
Environmental Research Letters 10: 084006.
http://iopscience.iop.org/article/10.1088/1748-9326/10/8/084006

Post E, Bhatt U, Bitz C, Brodie J, Fulton T, Hebblewhite M, Kerby J, Kutz S, Stirling I, Walker D. 2013. Ecological
consequences of sea-ice decline. Science 341: 519-524.
http://www.sciencemag.org/content/341/6145/519.full

Rubenstein, M. 2010.
http://blogs.ei.columbia.edu/2010/08/26/a-beginners-guide-to-climate-models/

TEACHER’S GUIDE
Grade level: High school
Key terms: models, climate change, extreme weather, sea ice loss, statistics, Arctic

3
 october 2015
The Arctic is melting - so what?

Glossary of Key Terms

Albedo effect − Derived from the Latin word for “whiteness”, the albedo effect is the amount of incoming
solar radiation that is reflected by a surface.

Arctic − An area surrounding the North Pole that includes the Arctic Ocean, most of Greenland, parts of
northern Russia, Alaska, Norway, Sweden and Finland, and most of the islands north of the Canadian mainland.
Not to be confused with the Antarctic which is area surrounding the continent of Antarctica at the South Pole.

Atmosphere − The envelope of gases surrounding Earth.

Climate − Weather conditions prevailing over a specific regions over a long period.

Equator − An imaginary line (at 0o latitude) around the Earth that is equidistant from the North and South
Poles, dividing the world into the Northern and Southern Hemispheres.

Evaporation − The changing of a liquid to a gas, often under the influence of heat (such as boiling of water).

Feedback − The modification of a process by its own results. For example, sea ice loss can “feedback” on
climate in two important ways: (1) through the “albedo” effect, and (ii) by lifting the “lid” on the heat stored in
the Arctic Ocean. The albedo effect refers to the reflection of incoming solar radiation by light-colored (white)
surfaces. If you have ever gone skiing on a sunny day, you know that the snow reflects visible light. It also
reflects infrared (heat) radiation. If there is less sea ice, then there is less solar reflection, which leads to more
heat being absorbed by the dark ocean water and continued sea ice loss. Sea ice can also cool the Arctic air
temperatures by insulating the heat within the Arctic Ocean. The Arctic Ocean never drops more than a couple
degrees below 0o C (or 32o F) (the dissolved salt in the water lowers the freezing point of ocean water). If the
ocean were exposed to air, it would warm Arctic air which is often below -20oC (-4o F) in winter.

Glacier − A persistent body of ice on land, formed by the accumulation of snow (and thus it is freshwater ice).
About two-thirds of the freshwater on Earth is in glaciers. If all glaciers were to melt, this would increase sea
level because this ice is not currently in the ocean.

Iceberg − a large (greater than 5 m across) floating mass of freshwater ice that has broken from a glacier or
ice shelf and is floating freely in open water. The saying “tip of the iceberg” comes from the fact that most of
the ice – over 90% - in an iceberg is below the surface of the water.

IPCC − The International Panel on Climate Change is a scientific body set up by the members of the United
Nations to explore the potential speed of climate change and its impacts.

Model − A representation of a thing or system. A paper airplane is a very simple model of a Boeing 747. The
simpler the model, the fewer the details that it incorporates.

Percentile − Given a set of numbers, the value below which a given percentage of observations fall. For
example, if we had the following set of numbers {4, 7, 3, 9, 3} the 60th percentile would be 4, because 3/5 x
100% = 60% of the set are equal to or less than 4.

4
 october 2015
The Arctic is melting - so what?

Glossary of Key Terms

Precipitation − Rain, snow, sleet or hail.

Sea ice − Frozen seawater. Sea ice covers about 7% of the Ocean surface at any given time. If all sea
ice were to melt now, it would have little impact on sea level rise because the ice formed from sea water
originally. However, it would impact sea level through climate “feedbacks” (decreased Arctic albedo and
Ocean heat insulation).
By the way, there really are about 200 terms for ice and they are important for scientists (compare the
implications of sea ice versus glaciers melting). The Editor’s favorite is “frazil”, the soft or amorphous ice
formed by the accumulation of ice crystals in water that is too turbulent to freeze solid. Imagine the ice
formed in the ocean under wave action! (See the following website for a description: http://www.arctic.
noaa.gov/essay_wadhams.html).

Statistically significant − A result that is likely not due to chance, but rather due to a real process.
Typically scientists define a result as “significant” if it would happen by chance less than 5% of the time.
The p-value is the scientists best estimate of how likely the result would be due to chance. Statistical
significance is an important way for scientists to deal with uncertainty. A new study in psychology finds that
p-values are doing their job: the lower the p-value the more likely the study is to be repeatable:
http://fivethirtyeight.com/datalab/psychology-is-starting-to-deal-with-its-replication-
problem/

Validation − to check the accuracy of something. For example, to validate a climate model, scientists run
the model to “predict” the weather patterns over the past 20 years. Then they compare the output with the
actual measurements and observations we have collected for that time.

Variance − The variance of a set of numbers is a measure of how different the numbers in the set are
from one another. For example, the two sets of numbers {6, 2, 3, 1} and {2, 4, 3, 3} have the same
average, or mean, of 3, but the first set has greater variance. The variance is calculated as:

where n is the number of x observations in the set, i is an index for each of the observations, Σ means take
the sum over the n observations in the set, and we assume that the mean, x ̅, is known ahead of time.
(If the mean is not known ahead of time, then the 1/n before the summation becomes 1/(n-1). If this
doesn’t make sense see Bessel’s correction in Wikipedia: https://en.wikipedia.org/wiki/Bessel%27s_
correction If it still doesn’t make sense, rest assured that this is college-level statistics.) For example, the
variance of the first set is [(6-3)2 + (2-3)2 +(3-3)2 +(1-3)2]/4 = 3.5. Prove to yourself that the variance of
the second set is 0.5.

Weather − The short-term conditions (temperature, dryness, windiness, etc.) of the atmosphere at a place
and time.